• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Radeon FURY thread

Soldato
Joined
4 Feb 2006
Posts
3,195
Very impressive performance from the Fury Pro.

The Witcher 3 @2160p looks a bit choppy though after playing it on a pile of TXs.:eek:

Well the Fury does holds up very well against a 980 clocked 384MhZ faster @ 1517MhZ. This just shows that clock for clock the Fiji architecture is faster. Just imaging a Fury @ 1500MHz :eek:. Let's hope voltage unlocking appears soon.
 
Associate
Joined
22 Feb 2015
Posts
445
Location
Dudley, UK
Another test showing AMD's better multi-GPU scaling, especially at higher resolutions where AMD's CPU overhead isnt as large. Not even just with the crazy 3/4 GPU builds too! 2 Fury X is faster than 2 Titan X at 4K in the majority of tests. In 3/4 GPU builds NVIDIA only wins in 3/13 tests, being BF4 and Crysis 3 and Thief.
 
Man of Honour
Joined
21 May 2012
Posts
31,940
Location
Dalek flagship
JQjBHBK.jpg

KjMMzGj.jpg

I could go on, but it's total domination for Fury X. :D

Now imagine drivers + voltage.

QSg8mRu.gif

I read this earlier and it is quite misleading.:D

Remember some of these cards are fantastic overclockers and they can also use voltage control.

If you crank up the settings at 2160p there is only going to be one winner as the 4gb and 6gb cards can not handle it 4 up.:)
 
Man of Honour
Joined
21 May 2012
Posts
31,940
Location
Dalek flagship
The settings are cranked. The 4GB isn't an issue.

That review is total garbage and 4gb is an issue @2160p

Here is proof that they were not using max settings, compare these results from the SEV3 bench to what they got lol.

2160p

1 GPU

  1. Score 25.8 GPU Fury X @1125/625, CPU 5960X @4.0, AMDMatt Link Mantle 15.15 Drivers
  2. Score 25.3 GPU 980 Ti @1350/1850, CPU 5930k @3.5, Clov!s Link DX11 353.30 Drivers
  3. Score 21.6 GPU 290X @1250/1625, CPU 3970X @4.9, AMDMatt Link Mantle 14.9 Drivers
  4. Score 21.2 GPU 290X @1270/1625, CPU 4930k @4.0, Kaapstad Link Mantle 14.9 Drivers
  5. Score 20.8 GPU 290X @1250/1625, CPU 3970X @4.9, AMDMatt Link DX11 14.9 Drivers
  6. Score 20.6 GPU 290X @1280/1625, CPU 4930k @4.0, Kaapstad Link DX11 14.9 Drivers
  7. Score 20.5 GPU 290P @1235/1625, CPU 3970X @4.9, AMDMatt Link Mantle 14.9 Drivers
  8. Score 18.5 GPU 980 @1502/2000, CPU 5960X @4.0, Kaapstad Link DX11 344.75 Drivers
  9. Score 17.1 GPU 295X2 @1018/1250, CPU 5960X @4.0, AMDMatt Link Mantle 14.12 Drivers
  10. Score 16.5 GPU 295X2 @1018/1250, CPU 5960X @4.0, AMDMatt Link DX11 14.12 Drivers

2 GPUs

  1. Score 40.3 GPU 290X @1240/1625, CPU 3970X @4.9, AMDMatt Link DX11 14.9 Drivers
  2. Score 39.8 GPU 290X @1230/1625, CPU 4930k @4.8, Kaapstad Link DX11 14.9 Drivers
  3. Score 32.1 GPU 295X2 @1018/1250, CPU 5960X @4.0, AMDMatt Link DX11 14.12 Drivers
  4. Score 27.4 GPU 980 @1550/2053, CPU 3930k @5.0, Besty Link DX11 344.16 Drivers
  5. Score 19.9 GPU 7990 @1175/1700, CPU 5960X @4.9, AMDMatt Link DX11 15.43 Drivers

3 GPUs




4 GPUs

  1. Score 99.1 GPU TitanX @1452/2002, CPU 5960X @4.5, Kaapstad Link DX11 353.06 Drivers
  2. Score 78.1 GPU 290X @1225/1625, CPU 5960X @4.4, AMDMatt Link DX11 14.11.2 Drivers
  3. Score 74.6 GPU 295X2 @1215/1625, CPU 5960X @4.9, AMDMatt Link DX11 14.12 Drivers
  4. Score 66.1 GPU 980 @1472/2002, CPU 5960X @4.0, Kaapstad Link DX11 347.09 Drivers
  5. Score 54.8 GPU 980 @1530/2000, CPU 3930k @5.0, Besty Link DX11 344.16 Drivers
  6. Score 48.3 GPU nvTitan @981/1752, CPU 3930k @4.0, Kaapstad Link DX11 344.11 Drivers


Another one they really screwed up was the Thief 2160p results, it is one of the few were the Fury X should beat the NVidia cards lol.:D
 
Soldato
Joined
22 Sep 2011
Posts
10,575
Location
Portsmouth (Southsea)
To be fair as mentioned above, stock vs stock is pointless - as it's starting to look like the Fury-X will not be great at over-clocking (namely the dev currently coding it into MSI afterburner), that lead may diminish.

That's just listing the specs. He might have OC'd, but I can't read Korean so gleaning any of those kind of details is impossible.
Not really, if you look at the single card specs at the bottom the Fury X is within a percentage of the 980ti.

If you factor in overclock/boost then it's clearly not the case as per every single other review comparing the two cards (the fury X falls behind significantly to an overclocked Titan/980ti).
 

bru

bru

Soldato
Joined
21 Oct 2002
Posts
7,360
Location
kent
So they got 71.9 fps in Sniper elite 3 on a single fury at 4k and our very own matt only managed 25.8 in the benchmark thread. (unless I'm reading that wrong, which is quite possible :))

Obviously something is not right there.


And only using a stock CPU is ludicrous for 4 card 4K benching. ( just like that other video)

I'm not saying that the Fury series do not scale very well in multi GPU setups but it looks so far as if the test setups have not been maxed out like we would run things and more like setup to give certain results.
 
Man of Honour
Joined
21 May 2012
Posts
31,940
Location
Dalek flagship
That's just listing the specs. He might have OC'd, but I can't read Korean so gleaning any of those kind of details is impossible.

Look closely at the scores and it is obvious that he did not OC the CPU, there are some terrible bottlenecks. Here is another one that demonstrates both a CPU bottleneck and the fact he was not using max settings.

Hitman scales very well using 1,2,3 or 4 cards @2160p yet look what happened when he ran the bench

7tZvgXV.jpg

He got the same scores for 2,3 and 4 cards lol.

I just ran Hitman on my 4 EVGA SC TXs which even at stock are 11% faster than reference TXs and this is what I got.

OzfYOR5.jpg

He was not using anywhere near max settings for Hitman.
 
Man of Honour
Joined
21 May 2012
Posts
31,940
Location
Dalek flagship
I'm not saying that the Fury series do not scale very well in multi GPU setups but it looks so far as if the test setups have not been maxed out like we would run things and more like setup to give certain results.

I think the Fury X is a very good card and once I have my 4 up and running with waterblocks I will max them out and test properly with max settings.
 
Back
Top Bottom